Communication Dans Un Congrès Année : 2025

Manipulating decision trajectories by altering task topology: empirical mouse-tracking studies and dynamical system model

Résumé

Background: In most empirical studies involving human decisions, self-reports as well as indirect measures such as response times and mouse-tracking usually rely on the user taking a decision among a fixed set of responses (e.g., two-alternative forced choice tasks, Likert scales). The psychometric impact of spatial and sensorimotor characteristics of the task are often neglected.
Aim: Integrating conceptual and sensorimotor topologies in the modeling of decision-making, by estimating the influence of various factors that could alter such topologies and therefore response dynamics and distributions.
Methods: Several empirical and computational studies were carried out to model decision-making processes reflected through mouse-tracking trajectories. The studies independently manipulated the spatial binding of responses on the computer screen by either grouping them (e.g., implicit association tasks), by manipulating target categories, or by allowing intermediate responses to prevent dichotomization in the decision process (e.g., continuous response scale). Other studies manipulated the sensorimotor cost to reach responses, relying on the dimensionality of the response space (2D vs. 1D) or biasing the movements of the mouse (e.g., speeding up movement towards an arbitrary response).
Results: Manipulated factors influence not only decision-making dynamics (associated to indirect measures) but also final response distributions (direct measures). These factors and their effects were mirrored in a computational model of decision-making based on differential equations operating on topologies (spatial, sensorimotor or conceptual).
Conclusion: Decision-making dynamics can be modelled as trajectories over a complex topology parametrized by the task design choices (e.g., type of response scale, target categories, their spatial configuration and accessibility).

Fichier non déposé

Dates et versions

hal-04975750 , version 1 (04-03-2025)

Identifiants

  • HAL Id : hal-04975750 , version 1

Citer

Jean-Charles Quinton, Flora Gautheron, Annique Smeding. Manipulating decision trajectories by altering task topology: empirical mouse-tracking studies and dynamical system model. International Conference on Spatial Cognition, Sep 2024, Rome, Italy. ⟨hal-04975750⟩
121 Consultations
0 Téléchargements

Partager

  • More